Experimental burns were effected and evaluated on a special variable temperature microscope stage designed specifically for this study. The cheekpouch of the golden hamster was everted, pinned across the viewing window and prepared for observation by removal of overlaying avascular tissue. A standard burn was created in all specimens by heating at a constant rate of 10 degrees K/sec to 56 degrees C, holding that temperature for 30 seconds, and returning to physiological temperature at a slower rate of -3 degrees K/sec. This protocol was preprogrammed and effected automatically by a special thermal regulatory unit designed to interface directly with the microscope stage. It produced a rather severe burn in the 2nd or 3rd degree class. Data indicate the effects of post burn cooling temperature and time duration of cooling. Several trends are clearly apparent: (a) When all other parameters are held constant, an optimum temperature for cooling is defined, above and below which recovery decreases; (b) The bandwidth of the recovery curve falls across a relatively narrow range of temperatures; (c) With increasing duration of cooling (1) the maximum recovery increases, (2) the range of effective temperature increases, and (3) the optimal cooling temperature increases; (d) No benefit is realized unless the tissue temperature is depressed to less than 20 degrees C. Reduction of the delay period between burning and the initiation of cooling results in a significantly greater recovery of circulatory function.